PDA

View Full Version : What Limit Switches?



Chaz
11-12-2016, 08:12 PM
So looking for both homing and limit switches.

This is for Thor which will have movement speed of at least 10m/min. Mechanical or Proxy? Will possibly using Index homing once sorted but I need to be able to stop a servo run away, so looking for something that will work and reliably.

audioandy
11-12-2016, 08:46 PM
So looking for both homing and limit switches.

This is for Thor which will have movement speed of at least 10m/min. Mechanical or Proxy? Will possibly using Index homing once sorted but I need to be able to stop a servo run away, so looking for something that will work and reliably.

Mechanical limit switches us the normally closed contacts.

Proxy for home switches.

That's how we do it at work anyway.

Andy

cropwell
11-12-2016, 10:20 PM
It took me a second or two to realise that proxy in this case did not mean one acting on behalf of another.
If you are using proximity switches mount them where they will not get covered in metal swarf.

Rob

Boyan Silyavski
12-12-2016, 12:03 AM
I always overdo but probably anyone will need only 4 switches, where 2 go on both side of long axis / If a router/ . At the end i mounted only 7 on my machine/ though bought 12 of them. But as Soft limits work, servo has its own limits, probably the board has its own software limits...maybe 4 will be enough.

I use Proximity, NC of course.

Another aspect which i overlooked is that they activate at 4mm, so basically to Home now i have to do it slowly so i will not overshoot, luckily the Chinese controller i use does up to 5 times homing, so if first time overshoots second time its precise.

What i am saying is i just found out that there are 8mm versions which may be better, so you could home faster, not that i am in a hurry. I think its worth checking and finding longer distance limit switches.



I start to think mechanical could be better idea, but they must be enclosed in solid case and avoid bending plate type. There should be clearly defined push axis and mount them respectively.



I was also mistaken to think that perpendicular to plate measured mounting will give me better precision. means i can hit the switch face and crush it if machine does not stop. / ok not so i i have hard stops, but you get the idea/. So if i am to do it again it will be side mounting where no interference will happen and switch will be better hidden on the machine, not bluntly visible.

cropwell
12-12-2016, 12:57 AM
I learnt the hard way that end-on proximity switches are vulnerable.

As far as homing accuracy is concerned Mach3 wallops up to the switch until it is triggered and then backs off until the switch is un-triggered, so I would imagine if you approach at warp speed, it would still maintain accuracy and consistency of the home position.

Rob

A_Camera
12-12-2016, 08:57 AM
So looking for both homing and limit switches.

This is for Thor which will have movement speed of at least 10m/min. Mechanical or Proxy? Will possibly using Index homing once sorted but I need to be able to stop a servo run away, so looking for something that will work and reliably.

I definitely vote for mechanical NC limit switches. The reason is that if something goes wrong with them the machine stops, you look for the error and eventually will find it. If you use proximity switches this may not always be the case, regardless if you use NC or NO type. Of course, for limit switches, nothing else but normally closed (NC) should be used, they guarantee broken wire detection also, which you can't have if you use normally open connection.

For home switches... well, I think mechanical is best but accurate mechanical switches are very expensive. I don't use any home switches, but have heard that proximity switches are accurate too. I guess it depends on the accuracy you need and the method of homing. Some software are better than others.

komatias
12-12-2016, 10:17 AM
Depends on the controller actually.

Mechanicals are fine for both limit and home but have you thought of maybe using the home on index function offered by some controllers? I have this on my GSK and I know that some of the CSMio's have this function too.

It would be the case of hitting a switch that initiates the slow movement and then does one full rotation of the screw until a flag is tripped or similar.

Chaz
12-12-2016, 10:56 AM
It took me a second or two to realise that proxy in this case did not mean one acting on behalf of another.
If you are using proximity switches mount them where they will not get covered in metal swarf.

Rob

Sorry, its an 'old' term we used to use as Electricians in South Africa.

Chaz
12-12-2016, 10:58 AM
So based on all of this, I am 'leaning' towards this type of limit switch. There is no 'end' sensor so if it runs part, it just pushes the arm.

Something like this - http://www.scoob.net/media/catalog/product/cache/1/image/7cc33aa79925eccc9980c5f8db9956ed/g/n/gnber_rme_8108-1.jpg

Chaz
12-12-2016, 12:20 PM
Any reason this will not be suitable?

This will be used to trigger into a CS Labs IP-A unit.

http://uk.rs-online.com/web/p/safety-limit-switches/6211515/

Neale
12-12-2016, 12:29 PM
I was talking to a friend recently who passed on some of his experience of using large numbers of mechanical switches in machine applications. He told me that the most common problem leading to unreliable triggering, lack of repeatability, etc, tended to be due to oxide build-up on the contacts. Two things you can do to help - use a higher signalling voltage (one reason why industrial machines use 24V rather than 5V) as this helps create microscopic arcing at the contacts which helps keep them clean, and also don't use them with high impedance input circuits. Drawing, say, 10mA or more through the contacts similarly helps them self-clean.

Personally, my machine is for home rather than commercial use so I'm only thinking of my own safety rather than meeting H&S legal requirements, and I've chosen to use proximity switches for combined home/limit and limit purposes. NC, of course, and mounted as pass-by rather than crash-into with the gap reduced to a minimum to help repeatability. Targets are long enough to ensure that the switch will remain triggered even if the moving part overruns. There is a failure mode in which they will fail in a non-safe mode (open collector output transistor goes short-circuit) but as they have built-in indicator LEDs, it is very easy to test them individually by just waving a small bit of metal near them and seeing if they trip.

I bought a box of 10 from eBay. So far, one seems to be faulty on arrival. One has a very slight but visible transition between on and off (judging by the LED brightness) but I've yet to test this with a real motion controller connection. The others seem to give a very sharp on/off transition and just turning the ballscrew by hand and watching the LED, I'm getting better than 1 div repeatability on a 0.01mm DTI. I expect that using the real motion controller will give even better results. This afternoon's job is completing the control box limit switch wiring...

Chaz
12-12-2016, 12:31 PM
I was talking to a friend recently who passed on some of his experience of using large numbers of mechanical switches in machine applications. He told me that the most common problem leading to unreliable triggering, lack of repeatability, etc, tended to be due to oxide build-up on the contacts. Two things you can do to help - use a higher signalling voltage (one reason why industrial machines use 24V rather than 5V) as this helps create microscopic arcing at the contacts which helps keep them clean, and also don't use them with high impedance input circuits. Drawing, say, 10mA or more through the contacts similarly helps them self-clean.

Personally, my machine is for home rather than commercial use so I'm only thinking of my own safety rather than meeting H&S legal requirements, and I've chosen to use proximity switches for combined home/limit and limit purposes. NC, of course, and mounted as pass-by rather than crash-into with the gap reduced to a minimum to help repeatability. Targets are long enough to ensure that the switch will remain triggered even if the moving part overruns. There is a failure mode in which they will fail in a non-safe mode (open collector output transistor goes short-circuit) but as they have built-in indicator LEDs, it is very easy to test them individually by just waving a small bit of metal near them and seeing if they trip.

I bought a box of 10 from eBay. So far, one seems to be faulty on arrival. One has a very slight but visible transition between on and off (judging by the LED brightness) but I've yet to test this with a real motion controller connection. The others seem to give a very sharp on/off transition and just turning the ballscrew by hand and watching the LED, I'm getting better than 1 div repeatability on a 0.01mm DTI. I expect that using the real motion controller will give even better results. This afternoon's job is completing the control box limit switch wiring...

Thanks. Good info. I'm using 24V, so this will help.

Can you link what you bought please?

Neale
12-12-2016, 12:36 PM
Thanks. Good info. I'm using 24V, so this will help.

Can you link what you bought please?

Sorry - it was sufficiently long ago that that seller doesn't seem to have them any more. They were "LJ12A3-4-Z/AX NPN NC 4mm Inductive Proximity Sensor Switch DC 3-wire 200mA", which are available from quite a few other suppliers. I paid Ģ23 for a box of 10 but again, prices vary and that was a couple of years ago.

Chaz
12-12-2016, 12:36 PM
Sorry - it was sufficiently long ago that that seller doesn't seem to have them any more. They were "LJ12A3-4-Z/AX NPN NC 4mm Inductive Proximity Sensor Switch DC 3-wire 200mA", which are available from quite a few other suppliers. I paid Ģ23 for a box of 10 but again, prices vary and that was a couple of years ago.

Thanks.

Chaz
12-12-2016, 12:38 PM
https://www.amazon.co.uk/LJ12A3-4-Z-3-wire-Cylindrical-Inductive-Proximity/dp/B0053FA8B8 Seems to be the same?

Neale
12-12-2016, 02:15 PM
Not absolutely sure. These specifically say "iron sensing" or somesuch. Because I don't have the original listing for mine, I can't say what they are. Proximity switches are sometimes iron only, sometimes anything metallic (although with reduced sensitivity, as I understand it). Mine certainly trigger on aluminium. Of course, if your trigger component is iron or steel, then that's not important.

Chaz
12-12-2016, 02:53 PM
Not absolutely sure. These specifically say "iron sensing" or somesuch. Because I don't have the original listing for mine, I can't say what they are. Proximity switches are sometimes iron only, sometimes anything metallic (although with reduced sensitivity, as I understand it). Mine certainly trigger on aluminium. Of course, if your trigger component is iron or steel, then that's not important.

Ive ordered 3 of them to test and see. They should be fine.

Thanks

Kardacian
13-12-2016, 12:19 PM
Ive ordered 3 of them to test and see. They should be fine.

Thanks
Would you consider using Hall Effect sensors? No moving parts extremely accurate and will outlast mechanical switches. Take a look at these, they run on 5 Volts but I can modify to use higher voltage if needed. They operate in the same manner as a NO switch. However a NC can also be created and will not be expensive.

https://www.tindie.com/products/Kardacian/magnetic-sensor-hall-effect/?pt=full_prod_search

Boyan Silyavski
13-12-2016, 03:11 PM
Would you consider using Hall Effect sensors? No moving parts extremely accurate and will outlast mechanical switches. Take a look at these, they run on 5 Volts but I can modify to use higher voltage if needed. They operate in the same manner as a NO switch. However a NC can also be created and will not be expensive.

https://www.tindie.com/products/Kardacian/magnetic-sensor-hall-effect/?pt=full_prod_search

Great but 5v and not shielded = Problems

Kardacian
13-12-2016, 04:10 PM
Great but 5v and not shielded = Problems


The voltage problem can be solved easily, what voltage do you need to support. I run these without shielding and have had no issues. What would be an ideal shielding for your environment?

m_c
13-12-2016, 05:02 PM
Great but 5v and not shielded = Problems
That's actually a bit of a generalisation, as it depends on exactly how you're powering things. I keep meaning to do a post on the why and how the problem can be mitigated to a certain extent.

Regardless, IMO limit switches should always be fail safe I.e. NC, as you want to minimise the risk of a wiring fault causing the switch detection to fail.
Plus, magnetic switches on something that could potentially be machining steel?

cropwell
13-12-2016, 05:30 PM
Great but 5v and not shielded = Problems biggest problem for me is the delivery charge !

Neale
13-12-2016, 05:44 PM
I see Hall effect switches mentioned more in US sites than UK. I'm not sure why you would use these rather than the Chinese-sourced proximity switches, unless you are in an environment where there is a lot of metal around and powerful magnet operation is desirable to avoid false triggering. For example, the proximity switches run natively on 24V. Is it really that there is a strong "not from China" feeling in the US rather than any pure technical reason? Just curious - I've a box of proximity switches in the workshop that I am using, so I've already made my design decision:smile:

JAZZCNC
13-12-2016, 07:48 PM
No Brainer to me Chaz.? . . . .The IP-A gives abilty to Home To Index so use it because will be superior to most SW and very repeatable.
Also means don't need expensive switches for there repeatabilty/reliabilty. Any decent Mechanical SW will work because all you doing is triggering the point Servo searches for the Index pulse.

Chaz
13-12-2016, 07:52 PM
No Brainer to me Chaz.? . . . .The IP-A gives abilty to Home To Index so use it because will be superior to most SW and very repeatable.
Also means don't need expensive switches for there repeatabilty/reliabilty. Any decent Mechanical SW will work because all you doing is triggering the point Servo searches for the Index pulse.

Thanks. Ill do some testing on the weekend. Ive not used the Index feature before, so will give it a shot.

hanermo2
13-12-2016, 11:02 PM
I am not (yet) homing on servo index w. CSMIO-IP-S-lathe, but ..
based on everything, I expect it to work as well as everything else so far, with some (major) issues sometimes.

I use 1:3 belt drive on the servo spindle.
The CSMIO-IP-S, mach3, lathe, driver cannot understand a transmission, thus will home to next 1:3 limit switch from the encoder.
With "position on spindle" or something similar.
Nothing wrong with it, just makes it impossible to get C axis.

For the time being, fttb,
any overrun is dangerous, potentially really fast (0.05-0.5 secs to crash), so..
For now, I use switches similar to the pics at start of thread.

They look good, are import, and are chinese copies badly made of a somewhat-decent industrial switch from omron (siemens, et al).
My current mechanical switches are not accurate, and donīt repeat well.
Afaik, afai can see, they cannot really fail to "break" on limits, thus they are ok as low-cost limit switches that look good.

My ongoing path will be;..
maintain the mechanical switches, always.
Independent.
They wont catch any crashes, these will come from the servo signals/overload, but usually, I think, with already (some) damage done.
Idea is hw limits prevent system over-run in sw /programming errors.
Both exist.
Imo. Ime.
With Csmio.
With M3.

In very, very, very rare cases it is possible for the cp to move "wrong" - seemingly forever at lowish rate (might be feed, limited feed, just donīt know).
I donīt know why.

Tracking error / load vs commanded path;
Medium-term (3-30 months), I even hope to catch this potential crash-error problem with a tracker system from the servo hw signal in-pos or near-pos.
This means, when running a repeated program, if the load exceeds a given level, or any servo drive has too much (little) load at that point, the program stops.
This would mean perfect "lights-out" running with near zero risk.


Home:
Index from z/ ie encoder, or separate opticals.
Separate opticals are cheap, easy, very very accurate.
Hw optical switches can do 0-1-2 microns, for 20$ or so, plus bits.
But encoder opticals can do 0.4 um, for 0$, and much better reliability.

A_Camera
14-12-2016, 09:50 AM
Would you consider using Hall Effect sensors? No moving parts extremely accurate and will outlast mechanical switches. Take a look at these, they run on 5 Volts but I can modify to use higher voltage if needed. They operate in the same manner as a NO switch. However a NC can also be created and will not be expensive.

https://www.tindie.com/products/Kardacian/magnetic-sensor-hall-effect/?pt=full_prod_search

NO switches are not recommended as limit switch. I would never use them in a mill, maybe a 3D printer, but otherwise your switches look nice.